Using array technology to understand dynamics of ... · Using array technology to understand dynamics of echolocation in bats and toothed whales Laura N. Kloepper, Ph.D. Brown University

PostDoc Journal Journal of PostDoctoral ResearchVol. 1, No. 4, April 2013 www.postdocjournal.comUsing array technology to understand dynamics of echolocation in bats andtoothed whalesLaura N. Kloepper, Ph.D.Brown University Department of Neuroscience, Providence, RIUMass Dartmouth Department of Electrical and Computer Engineering, Dartmouth, MAemail: laura [email protected]

AbstractEcholocation, a unique sensory strategy of projecting and receiving ultrasonic signals to perceivethe environment, is a dynamic process used by bats and toothed whales that allows the emitter toadapt its sound to environment and echolocation task. This paper reviews the application ofacoustic array sensing to understand echolocation dynamics in Microchiroptera and Odontoceti.

Introduction

Echolocation, or biosonar, is a sensory modal-ity that evolved independently in both mi-crochiropteran bats (family Microchiroptera) andtoothed whales (suborder Odontoceti). Be-cause bats primarily forage during the nightand odontocetes often forage at depth of lim-ited light, these animals developed a uniquestrategy to the challenges of hunting for preyin an environment with very little ambient light.Instead of relying primarily on vision, these ani-mals emit short, intense, ultrasonic sounds anduse the reflected echoes to navigate and locateprey.

Figure 1: Photo of Carollia perspicillata, a nose-leaf bat. Sound is projected from the nose-leafstructure which helps create the shape of thesound beam. Photo credit: Mark Terk.

Microchiropteran bats produce their echoloca-tion sounds in their larynx and emit themthrough either their nose or highly developednoseleaf structures (Fig. 1). The signals aretypically 1 to 25 ms in duration, with frequen-cies between 20 and 110 kHz [37, 10, 20, 34].The echolocation calls of bats are species spe-cific and can be frequency modulated (FM), con-stant frequency (CF), or a combination of both.Odontocetes, on the other hand, produce theirsignals pneumatically using highly specializednasal structures [40, 31, 11, 22]. Some groupsof species produce FM sweeps or low frequencysignals, but the most of the odontocete speciesproduce short, high frequency signals (for pur-poses of this paper I will further only refer tothese groups, the Delphinid and Phocoenid fam-ilies of Odontoceti). Inside the forehead lie apair of structures called the monkey lips dorsalbursae (MLDB) complex. Each MLDB containsa valve-like structure composed of dense tissue(the phonic lips) and a pair of fat bodies (thedorsal bursae) [11]. To produce sound, odon-tocetes pass high pressure air across one orboth of the phonic lips which causes the lipsto slap together and create echolocation pulses[12, 26, 29, 13]. These pulses are directedthrough a fat-filled melon, and emitted into thewater from the anterior forehead (Fig. 2) [1, 3].These signals are typically 10 to 70 µs in dura-tion, with frequencies between 25 and 130 kHz[35]. Delphinids produce signals with most en-ergy in a range of approximately 80 kHz (termedbroadband), and phocoenids produce signalswith most energy in a range of approximately 20

2 Journal of Postdoctoral Research 2013: 1-8

kHz (termed narrowband) [2, 17, 5, 35, 46].

Figure 2: Schematic of the sound generationapparatus for a false killer whale. Green= MLDBcomplex, red= nasal plug muscle, purple= airsacs, blue= melon, beige= bone.

Directionality of Echolocation Signals

Despite differences in echolocation producingstructures and call type, these two groups ofmammals both produce signals that are direc-tional, meaning they do not emanate equallyfrom all directions of the animal and have ameasurable sound beam. Due to the natureof sound, higher frequency sounds will havea greater directionality (or a narrower soundbeam) than lower frequency sounds [19]. Thisdirectionality is key to the design of the echolo-cation system: by producing signals with a nar-row range of ensonification, the animal canincrease the amount of energy reflected backfrom a target of interest and decrease the en-ergy reflected back from surrounding clutter.

Regardless of the initial directionality producedby the animal, all sound beams will spread asthey are transmitted through a medium suchthat the sound beam will cover a larger area asthe distance between the animal and the targetincreases. As such, a narrow sound beam willbe beneficial for reducing echoes from clutterwhile tracking prey at long ranges, but mightbe disadvantageous for tracking prey at narrowranges, especially if the prey is able to makeevasive maneuvers [39, 42, 41]. If the soundbeam is sufficiently narrow, any movement by a

prey item at close range might put the prey out-side the area of the sound beam and render itacoustically invisible to the echolocating animal.Therefore, there is an inherent tradeoff betweencreating a sound beam narrow enough to re-duce clutter at wide ranges yet broad enough totrack prey at close ranges. A strategy of dynam-ically changing the width of the echolocationbeam according to the distance of the prey tar-get or degree of clutter would counteract theseeffects.

Recently, research with both bats and odon-tocetes suggest that these animals might beable to do just this. These changes in beamangle, size or shape might be accomplished bymanipulating structures responsible for soundproduction and beam shape formation. As men-tioned above, depending on the species eitherthe mouth or the noseleaf serves as the acous-tic emitter for bats. Experiments demonstratethat changes in the mouth opening or noseleafshape can dramatically alter the emitted beamshape, so the bat can make rapid, fine scaleadjustments in the shape of these structures tomanipulate the directionality of the echolocationbeam [21, 49, 50, 45, 44]. With odontocetes, itis hypothesized that they use their melon, whichis composed of unique acoustic fats, to helpfocus the emitted sound [14]. The melon con-tains a topography of fats with variable soundspeed transmissions that direct sound towardsthe front and center of the melon, where it isthen transmitted into the water [27, 36, 30]. Thismelon is surrounded by a vast network of mus-cles that may act to change the curvature of themelon and change the directionality of the emit-ted beam much like an acoustic lens [18].

Acoustic Array Sensing Technology

One of the biggest challenges in any scien-tific field is developing research equipment andmethods that are reliable, precise, easy to repli-cate and affordable. Research with echolocat-ing animals is no exception. Animal vocaliza-tions are measured using microphones (in air)or hydrophones (in water). These devices are

Laura N. Kloepper 3

electroacoustic transducers that, on the sim-plest level, convert acoustic energy to electri-cal energy. Most microphones used for batresearch are condenser microphones, whichprovide an electrical signal as sound waves dis-place a diaphragm located inside the micro-phone case. An emergent sensor technologyknown as micro electro-mechanical systems(MEMS) are quickly becoming popular in thedesign of bioacoustics arrays [15]. Known bestfor their use in cellular phones, low-cost MEMSdevices are commercially available in a widevariety of frequency ranges and sensitivities.These MEMS devices have the advantage thatthey can be as effective as many larger micro-phones but are a fraction of their size. By mea-suring the tiny capacitance change between twointerleaved fingers, MEMS devices can producea voltage response to the sound pressure wavesreceived by the exposed silicon die.

Most hydrophones are composed of materi-als with piezoelectric or electrostriction proper-ties, meaning they gain an electric charge whenthey receive a physical stress (such as a soundwave impinging on the element). To measurethe sound at one location, only one receiver,or element, is needed; however, to capture theecholocation beam shape, multiple receiversare needed. This is accomplished with an arrayof elements in a known geometric configuration.Common array configurations include linear ar-rays, y-shaped arrays, or planar arrays and maybe densely or sparsely populated (Fig. 3).

Figure 3: Examples of common array configu-rations. A) linear array, B) y-shaped array, C)cross section planar array, D) radial planar ar-ray.

The process of recording and reconstructingthe beam shape is relatively simple. All of thecalibrated elements in the array simultaneouslyrecord each echolocation signal. Once the sig-nals have been recorded, the beam reconstruc-tion occurs via off-line analysis. For each signal,the element with the highest recorded amplitude(in units of db SPL re 20µPa for air and dB SPLre 1µPa for water) is considered the center, or”on-axis” element, for that signal. The ampli-tudes are then calculated for the remaining el-ements of the array and the beam angle, size,or shape is calculated via interpolation betweenthe array elements.In order to quantify the beam, a criterium mustbe set for amplitude values. Typically, a -3dBlimit is set relative to the on-axis element anddefined as the main lobe emitted by the animal.For example, if the on-axis element has an am-plitude value of 180 dB, the -3dB beam would bethe elements (or interpolated value between theelements) where the amplitude is 177 dB. Be-cause the decibel scale is logarithmic, the -3dBportion of the beam represents the area withhalf the power of the on-axis signal. For lineararrays, this -3dB value is typically given in az-imuth and elevation angles assuming a symmet-ric beam shape. For planar arrays, these val-ues can be given in terms of beam area, which

4 Journal of Postdoctoral Research 2013: 1-8

allows for asymmetric beam shapes (Fig. 4).

Applying array sensing as a biological re-search tool

The basic structure and design of acoustic ar-rays allows for measurement of biological sig-nals under a variety of conditions in the lab-oratory and the field. For odontocetes, someof the most commonly used field array designsare linear and Y-shaped arrays (Fig. 3A and3B). These arrays can be stationary or towedoff boats to localize individuals and measurethe echolocation signals of free-ranging an-imals [32, 9, 28]. While field arrays providegreat insight into the foraging behavior of wildodontocetes, arrays used in laboratory settingsallow for fine scale measurements of the beamin a controlled environment. Perhaps the mostbasic information is that of beam shape. Al-though some species of odontocetes, such asthe false killer whale (Pseudorca crassidens)produce a single-lobed, symmetric beam [25],others, such as the bottlenose dolphin (Tursiopstruncatus) produce a beam with a dual-lobedstructure [43]. This is thought to aid in trackingprey and might be achieved by using both setsof phonic lips for sound production. The phoniclips of some odontocetes are asymmetric andit is thought that low-frequency signals are pro-duced by the larger right side, high frequencysignals are produced by the smaller left side,and the two signals are combined within themelon [7, 26, 13, 43]. Some bats, such as thebig brown bat (Eptesicus fuscus) also produce abeam with a dual-lobed structure, and althoughthe mechanism is unknown, it its thought to aidin altitude determination [16]. Instead of point-ing the main axis of their beam at their target,Egyptian fruit bats (Rousettus aegyptiacus) usea strategy of directing the maximum slope oftheir beam at the target, which is likely optimizestarget localization [48]. Although this has not yetbeen directly tested in odontocetes, prior mea-surements of echolocation beam direction showsome species direct their echolocation beamslightly upwards or downwards which might in-dicate a utilization of slope for target detection.

For example, the -3dB echolocation beam hasbeen measured at an upward angle of 5 de-grees for the beluga whale (Delphinapterus leu-cas) [8], upwards angles of both 5 [6] and 20 de-grees for the bottlenose dolphin [4], 0 degreesfor the harbor porpoise [5], and a downward an-gle of 5 degrees for the false killer whale [7].These differences in beam direction might bea result of species specific differences in skullanatomy [11] or might be due to active beamsteering during echolocation [33].

Figure 4: Recreation of the echolocation beamvia the interpolation method for a radial planararray. The color bar represents the source level,or intensity of the signal interpolated across thearea sampled by the array. The black line in-dicates the -3dB contour of the beam and thebeam area can be calculated from this shape.

In addition to understanding the basic shapeof the echolocation beam, array measurementsallow for a more intricate investigation into howthe beam changes during different echoloca-tion scenarios. For example, Daubentons bats(Myotis daubentonii) have been shown to nar-row their echolocation beam while searchingfor insects in the field versus in the laboratory[44]. This increase in directionality occurs witha concomitant increase in signal intensity andmight be accomplished by increasing the widthof the mouth opening during echolocation. Pro-ducing louder, more directional signals has abenefit for bats flying in the field: more en-ergy is concentrated in a narrower direction,which results in stronger echoes from targets

Laura N. Kloepper 5

of interest and quieter echoes from surround-ing clutter. Such a strategy is likely employedby odontocetes as well, as echolocation sig-nals from free-ranging individuals tend to havehigher directionality than laboratory individuals[47, 38]. Adjustment of beam width is not limitedsolely to echolocation environment, and recentstudies indicate that both bats and odontocetescan change their beam depending on the tar-get distance or direction. Both Daubentonsbats and Serotine bats (Eptesicus serotinus)widen their echolocation beams during the ter-minal phase of echolocation [23]. Widening thebeam when prey are close allows bats to com-pensate for the disadvantage of a directionalbeam and would increase the probability of preydetection at close range. False killer whaleschange the size of their beam depending on tar-get characteristics or distance which might bea strategy of reducing the size of the beam tomaximize the energy in the reflected echo [24].Perhaps even more remarkable is evidence thatbottlenose dolphins can widen and even steertheir echolocation beam depending on angulartarget location, which further suggests internalbeam control via structures associated with themelon [33]. With the array studies to date, weare only beginning to get a glimpse of the capa-bility of the fine acoustic adjustments performedby these animals during echolocation. As com-puting power increases and cost of equipmentdecreases, the amount of information availablefrom array technology will become more de-tailed and allow for further in depth investigationinto the dynamics of echolocation.

AcknowledgementsThe author would like to thank Jason Gaudette,James Simmons, and two anonymous review-ers for valuable comments on the manuscript.This was partially funded by the National Sci-ence Foundation Postdoctoral Research Fellow-ship in Biology DBI-1202833 awarded to LauraN. Kloepper.

References

[1] J L Aroyan, T W Cranford, J Kent, andK S Norris. Computer Modeling of Acous-tic Beam Formation in Delphinus Delphis.J Acoust Soc Am, 92:2539–2545, October1992.

[2] W W Au, R W Floyd, R H Penner, and A EMurchison. Measurement of echolocationsignals of the Atlantic bottlenose dolphin,Tursiops truncatus Montagu, in open wa-ters. J Acoust Soc Am, 56:1280–1290, Au-gust 1974.

[3] W W L Au. Acoustic radiation fromthe head of echolocating harbor por-poises (Phocoena phocoena). J Exp Biol,209(14):2726–2733, July 2006.

[4] W W L Au, R W Floyd, and J E Haun.Propagation of Atlantic bottlenose dolphinecholocation signals. J Acoust Soc Am,pages 1–12, August 1978.

[5] W W L Au, R A Kastelein, T Rippe, and N MSchooneman. Transmission beam patternand echolocation signals of a harbor por-poise (Phocoena phocoena). J Acoust SocAm, pages 1–7, November 1999.

[6] W W L Au and P W Moore. Echoloca-tion transmitting beam of the Atlantic bot-tlenose dolphin. J Acoust Soc Am, pages1–4, October 1986.

[7] W W L Au, D Pawloski, P E Nachtigall,M Blonz, and R C Gisner. Echolocationsignals and transmission beam pattern of afalse killer whale (Pseudorca crassidens).J Acoust Soc Am, pages 1–9, October1995.

[8] W W L Au, R H Penner, and C W Turl.Propagation of beluga echolocation sig-nals. J Acoust Soc Am, pages 1–7, Oc-tober 1987.

[9] Whitlow W L Au and Kelly J Benoit-Bird. Automatic gain control in the

6 Journal of Postdoctoral Research 2013: 1-8

echolocation system of dolphins. Nature,423(6942):861–863, June 2003.

[10] Murat Aytekin, Beatrice Mao, and Cyn-thia F Moss. Spatial perception and adap-tive sonar behavior. J Acoust Soc Am,128(6):3788–3798, 2010.

[11] T Cranford, M Amundin, and K S Norris.Functional Morphology and Homology inthe Odontocete Nasal Complex: Implica-tions for Sound Generation. Journal ofMorphology, pages 1–63, October 1996.

[12] T W Cranford and M Amundin. Biosonarpulse production in odontocetes: the stateof our knowledge. In J A Thomas, C FMoss, and M Vater, editors, Echolocationin Bats and Dolphins, pages 27–35. Uni-versity of Chicago Press, Chicago, Febru-ary 2004.

[13] Ted W Cranford, Wesley R Elsberry,William G Van Bonn, Jennifer A Jeffress,Monica S Chaplin, Diane J Blackwood,Donald A Carder, Tricia Kamolnick, Mark ATodd, and Sam H Ridgway. Observationand analysis of sonar signal generationin the bottlenose dolphin (Tursiops trunca-tus): Evidence for two sonar sources. Jour-nal of Experimental Marine Biology andEcology, 407(1):81–96, October 2011.

[14] W E Evans, W W Sutherland, and R GBeil. The directional characteristics of Del-phinid sounds. Marine Biology, pages 1–21, November 1964.

[15] J E Gaudette, L N Kloepper, and J A Sim-mons. Large reconfigurable microphonearray for transmit beam pattern measure-ments of echolocating bats. J Acoust SocAm, 131(4):3361, March 2012.

[16] Kaushik Ghose, Cynthia F Moss, and Tim-othy K Horiuchi. Flying big brown bats emita beam with two lobes in the vertical plane.J Acoust Soc Am, 122(6):3717, 2007.

[17] A D Goodson and C R Sturtivant. Sonarcharacteristics of the harbour porpoise

(Phocoena phocoena): source levels andspectrum. ICES Journal of Marine Sci-ence, 53:465–472, April 1996.

[18] C J Harper, W A Mclellan, S A Rommel,D M Gay, R M Dillaman, and D A Pabst.Morphology of the melon and its tendi-nous connections to the facial muscles inbottlenose dolphins (Tursiops truncatus).Journal of Morphology, 269(7):820–839,2008.

[19] G R Harris. Review of transient field the-ory for baffled planar piston. Journal of theAcoustical Society of America, 70:10–20,October 1981.

[20] S Hiryu, M E Bates, J A Simmons,and H Riquimaroux. FM echolocatingbats shift frequencies to avoid broadcast-echo ambiguity in clutter. Proceedingsof the National Academy of Sciences,107(15):7048–7053, April 2010.

[21] Shizuko Hiryu, Koji Katsura, Liang-KongLin, Hiroshi Riquimaroux, and YoshiakiWatanabe. Radiation pattern of echolo-cation pulse in Taiwanese leaf-nosed bat,Hipposideros terasensis. Acoust. Sci. &Tech., 27(2):108–110, 2006.

[22] D S Houser. Structural and functionalimaging of bottlenose dolphin (Tursiopstruncatus) cranial anatomy. J Exp Biol,207(21):3657–3665, October 2004.

[23] L Jakobsen and A Surlykke. Vespertil-ionid bats control the width of their biosonarsound beam dynamically during prey pur-suit. Proceedings of the National Academyof Sciences, 107(31):13930–13935, Au-gust 2010.

[24] L N Kloepper, P E Nachtigall, M J Don-ahue, and M Breese. Active echolo-cation beam focusing in the false killerwhale, Pseudorca crassidens. J Exp Biol,215(8):1306–1312, March 2012.

[25] Laura N Kloepper, Paul E Nachtigall,Christopher Quintos, and Stephanie A Vla-

Laura N. Kloepper 7

chos. Single-lobed frequency-dependentbeam shape in an echolocating false killerwhale (Pseudorca crassidens). J AcoustSoc Am, 131(1):577, 2012.

[26] M O Lammers and M Castellote. The bel-uga whale produces two pulses to form itssonar signal. Biology Letters, 5(3):297–301, May 2009.

[27] C Litchfield and A J Greenberg. Com-parative lipid patterns in the melon fats ofdolphins, porpoises and toothed whales.Comparative Biochemistry and physiology,B., 47:1–8, March 1973.

[28] P T Madsen. Echolocation clicks of twofree-ranging, oceanic delphinids with differ-ent food preferences: false killer whalesPseudorca crassidens and Risso’s dol-phins Grampus griseus. J Exp Biol,207(11):1811–1823, May 2004.

[29] P T Madsen, D Wisniewska, and K Beed-holm. Single source sound production anddynamic beam formation in echolocatingharbour porpoises (Phocoena phocoena).J Exp Biol, 213(18):3105–3110, August2010.

[30] D C Malins and U Varanasi. CetaceanBiosonar Part II: The Biochemistry of lipidsin acoustic tissues. In D C Malins andJ R Sargent, editors, Biochemical and Bio-physical Perspectives in Marine Biology,pages 237–287. Academic Press, London,November 1975.

[31] J G Mead. Anatomy of the External NasalPassages and Facial Complex in the Del-phinidae (Mammalia: Cetacea). Smithso-nian Contributions to Zoology, pages 1–80,November 1975.

[32] B Mohl, M Wahlberg, P T Madsen, L AMiller, and A Surlykke. Sperm whale clicks:Directionality and source level revisited. JAcoust Soc Am, pages 1–11, December2000.

[33] Patrick W Moore, Lois A Dankiewicz, andDorian S Houser. Beamwidth control andangular target detection in an echolocatingbottlenose dolphin (Tursiops truncatus). JAcoust Soc Am, 124(5):3324, 2008.

[34] C F Moss and A Surlykke. Probing the nat-ural scene by echolocation in bats. Front.Behav. Neurosci., 4:1–16, 2010.

[35] K Nakamura and T Akamatsu. Compari-son of Click Characteristics among Odon-tocete Species. In J A Thomas, C F Moss,and M Vater, editors, Echolocation in Batsand Dolphins, pages 36–40. University ofChicago Press, Chicago, February 2004.

[36] K S Norris and Harvey, G W. Sound Trans-mission in the Porpoise Head. J AcoustSoc Am, 56:659–664, October 1974.

[37] Anthony E Petrites, Oliver S Eng, Don-ald S Mowlds, James A Simmons, andCaroline M DeLong. Interpulse intervalmodulation by echolocating big brown bats(Eptesicus fuscus) in different densitiesof obstacle clutter. J Comp Physiol A,195(6):603–617, March 2009.

[38] Marianne H Rasmussen, MagnusWahlberg, and Lee A Miller. Esti-mated transmission beam pattern of clicksrecorded from free-ranging white-beakeddolphins (Lagenorhynchus albirostris). JAcoust Soc Am, 116(3):1826, 2004.

[39] K D Roeder. Acoustic sensory responsesand possible bat-evasion tactics of certainmoths. In Proceedings of the Canadian So-ciety of Zoologists Annual Meeting, pages71–78, February 1974.

[40] E J Schenkkan. On the comparativeanatomy and function of the nasal tract inodontocetes (Mammalia, Cetacea). BijdrDierk, 43:127–159, November 1973.

[41] N Skals and A Surlykke. Hearing and eva-sive behaviour in the greater wax moth,

8 Journal of Postdoctoral Research 2013: 1-8

Galleria mellonella (Pyralidae). Physio-logical Entomology, 25:354–362, February2000.

[42] H G Spangler. Moth hearing, defense, andcommunication. Annual Reviews of Ento-mology, 33:59–81, July 1988.

[43] J Starkhammar, P W Moore, L Talmadge,and D S Houser. Frequency-dependentvariation in the two-dimensional beam pat-tern of an echolocating dolphin. BiologyLetters, 7(6):836–839, November 2011.

[44] A Surlykke, S Boel Pedersen, and L Jakob-sen. Echolocating bats emit a highly direc-tional sonar sound beam in the field. Pro-ceedings of the Royal Society B: Biologi-cal Sciences, 276(1658):853–860, March2009.

[45] Dieter Vanderelst, Fons De Mey, HerbertPeremans, Inga Geipel, Elisabeth Kalko,and Uwe Firzlaff. What Noseleaves Dofor FM Bats Depends on Their Degreeof Sensorial Specialization. PLoS ONE,5(8):e11893, August 2010.

[46] A Villadsgaard, M Wahlberg, andJ Tougaard. Echolocation signals of

wild harbour porpoises, Phocoena pho-coena. J Exp Biol, 210(1):56–64, January2007.

[47] Magnus Wahlberg, Frants H Jensen, Nat-acha Aguilar Soto, Kristian Beedholm, LarsBejder, Claudia Oliveira, Marianne Ras-mussen, Malene Simon, Anne Villads-gaard, and Peter T Madsen. Source pa-rameters of echolocation clicks from wildbottlenose dolphins (Tursiops aduncus andTursiops truncatus). J Acoust Soc Am,130(4):2263–2274, 2011.

[48] Y Yovel, B Falk, C F Moss, andN Ulanovsky. Optimal Localiza-tion by Pointing Off Axis. Science,327(5966):701–704, February 2010.

[49] Qiao Zhuang and Rolf Muller. Nose-leaf Furrows in a Horseshoe Bat Act asResonance Cavities Shaping the BiosonarBeam. Phys. Rev. Lett., 97(21):218701,November 2006.

[50] Qiao Zhuang and Rolf Muller. Numer-ical study of the effect of the noseleafon biosonar beamforming in a horseshoebat. Phys. Rev. E, 76(5):051902, Novem-ber 2007.